Codebase list freetds / debian/0.63-3 TODO
debian/0.63-3

Tree @debian/0.63-3 (Download .tar.gz)

TODO @debian/0.63-3raw · history · blame

This list is ordered top-to-bottom by priority.  
Things that are broken and need mending are at the top, 
followed by things that should work before the next release, 
followed by features that should be added/fixed/reworked (grouped by library).  

Everyone is encouraged to add to the list.  Developers can do it directly; 
anyone else can post a patch to Source Forge.  
In this way we can communicate with each
other about the project's priorities and needs.  

To Do List	$Id: TODO,v 1.117 2004/11/04 13:11:32 jklowden Exp $
------------
Broken:
. Undelimited data files should be read w/ tds_iconv_fread().
. Passing invalid character set names to server with tsql?
. Update involving triggers may not set rows_affected correctly
  Cf. ML 08 Jan 2004 "New BUG: update/delete returns -1 instead of correct
      affected rows".  

Postponed to 0.63 version
. log and call error handler on problems with tds_iconv_init().
. ct_dynamic and friends needed for DBD::Sybase placeholder support
 (work in progress)
. iconv
  - support string conversions for Sybase
  - add test for locale_charset() to configure.in.  Use it if available
    instead of nl_langinfo, because it implies we're using GNU iconv
. add doc/htdoc/Makefile with "publish" target to commit cvs and update 
  freetds.org
. drop txt2man from cvs, skip building man pages without it. 
. Rename TDS 8.0 to TDS 7.1, to be consistent with Microsoft's name
. be able to disable iconv for BCP (see Sybase documentation)
. UTF-8 unit tests (libTDS)
  - very large column name (128 on mssql), no-english to UTF-8 client
. userguide
  - update table of working configurations
    server (vendor, version, platform) - client (freetds version, platform) 
    - protocol - charsets - date

Started in 0.63
- rationalize libtds structure and *_alloc functions 
  (ML 5 December 2002, sb:"names")
  (remove INFO and others)
- cancel. We added a test, analyzed the problem but no code... 
  We have two options: comment test or start coding.
  * test commented (document in README ??)
- ODBC data test fails for numeric, convert.c do not convert 
  NUMERIC ->NUMERIC changing precision/scale
  * this part never worked, commented
- limit data sended via input parameters/bulk copy
  (ie SYBCHAR to 255, XSYBCHAR to 8000)
  test behavior of proprietary library (warning or not)

For future versions (in priority order within library):
  All:
. move error message macro/prototype to tds.h.in; use in dblib.c.
. Remove old char(30) limitations to various sysnames e.g., tables, columns, 
  passwords (checks).  Is this done?  
. Add missing constants needed for python and verify working
  (work in progress)
. Add missing constants needed for Gnome-DB and verify working 
. conversion from ucs2 to utf8, provide for 2+ bytes/character
. tdsping program for testing purposes
. Finish off the TDS dissector for ethereal
. more solid. handle out of memory conditions (started, we must test all 
  allocation and all function that return allocated data and be able to 
  rollback to a safe state)
. optimize numeric conversions (partial)
. limit copy of data (now wire -> tds -> temp -> client)
  use a method like zero-copy
  In token.c getting row should write data even to client
  (callback, new user types, see ctlib conversion functions)
  In query.c writing data do the same (use additional column fields?)
. review the way parameters are packed 
  (too complicate, see ctlib bulk, cf "bulk copy and row buffer")
. cursor support (limited support on ctlib)
. tds_connect free tds structure on failure. This cause in some condition
  a pointer to garbage in some library (workaround written, best to change)
. support for named instances under mssql2k
. Autodetect TDS protocol when connecting (and cache somewhere?)
. support for VARIANT type (requested one time, 2003-8-1)
. support for NT named pipe (requested many time ago for mssql6.5 server, only
  for completeness)
. NTLMv2 domain login support.  Cf. ML 5 Aug 2003 12:20:21 -0700 (PDT) 
  from bradleyb@u.washington.edu re: "Domain Login Support, connecting to MSSQL issue"
  and http://davenport.sourceforge.net/ntlm.html.  For now, document.
. support encrypted connections
. read on partial packet, do not wait entire one
. net.c file with all network stuff
. detect if realloc/free accepts NULL pointers (in configure.in)
. support for password longer than 30 characters under Sybase
. optimization: add -Wl,-Bsymbolic or similar to libraries
. under Sybase using prepared statements and BLOBs we shouldn't try to
  prepare every time (cache failure preparing, see odbc unittests logs)

  db-lib
. Returning INT_CANCEL from an error handler doesn't always engender FAIL. 
. add DBTEXTLIMIT (dbsetopt), PHP require it to support textlimit ini value

  ct-lib
. dynamic placeholders (DBD::Sybase)
. ct_option() calls (CS_OPT_ROWCOUNT, CS_OPT_TEXTSIZE, among others)
. async function, async calls (dbpoll() and friends)
. support all type of bind in ct_bind (CS_VARBINARY_TYPE and other)
  search "site:.sybase.com CS_VARBINARY ct_bind" on google for more info
. complete sqlstate and other field in message (for Python)

  odbc
. Star Office complains that these TypeInfo constants are not implemented in SQLGetInfo:
	47      SQL_USER_NAME
  (handle environment callbacks)
  do a "SELECT USER_NAME()". If data pending MS do another connection with 
  same login.
. SQLNativeSql and fill SQLGetInfo according (mssql7+ handle odbc escapes 
  directly)
. autoinstall for RPM
. constants parameters in call syntax (ie {?=func(10,?)} )
. SQLDescribeParam (Sybase seems to require it)
. change ODBC_RETURN to return errs.lastrc (not to set) and use errs.lastrc
  to store error, cache some errors (truncated output)
  optimize ODBC_RETURN (remove useless)
. handle no termination on odbc_set_string*

. test and fix: select with compute
. test and fix: hidden fields (FOR BROWSE select)
. test: all binding types (input and output)
. test: new descriptors work
  - ODBC 2 type returned (datetime)
  - SQLDescribeCol return ODBC 2 type or 3 or based on configuration?
  - what happen to SQL_DESC_DATETIME_INTERVAL_CODE and SQL_DESC_CONCISE_TYPE
    changing only SQL_DESC_TYPE (with SQLSetDescField)
. test: set SQL_C_DEFAULT and call SQLFetch (numeric, others)
. test: SQLGetStmtAttr(SQL_ATTR_ROW_NUMBER)
  - all binded parameters
  - no bind and sqlgetdata
  - before first fetch
  - after last fetch
. bind offset + tests
. PARAMSET_SIZE and related


  pool
. get connection pooling working with all protocol versions
. Make pool configuration a non-recompile process

  server
. Server API needs more work, especially for TDS 5.0/7.0

Done
. FAQ
  - multiple statements question
  - PHP workaround with multiple connections (see ML)
. ODBC test: SQLPutData/SQLGetData
. ODBC: see putdata test SQLPutData really broken
. DBD::Sybase uses the ct_command(CS_RPC_CMD) and ct_param()
. cs_calc and cs_cmp, for sybperl
. ct_get/send_data()
. Array binding for ctlib not working
. bcp support to ctlib
. support ct_command/CS_SEND_DATA_CMD, now crash
  (see http://manuals.sybase.com/onlinebooks/group-cn/cng1000e/ref/@Generic__BookTextView/50456)